A Proficient Convolutional Neural Network for Classification of Bone Age from X-Ray Images

Bone age evaluation is crucial for identifying and planning interventions for numerous disorders.Estimating bone age is distinct from assessing physical development based on an individual's birth date.This evaluation of bone age reveals growth and progression, facilitating the identification and management of pediatric diseases.Significant obstacles in bone age evaluation often stem from low-quality X-ray images, obscured bone formations, and the intricacies of feature extraction due to compromised image quality, which greatly affects the performance of models.This research introduces VGG19, a groundbreaking Convolutional Neural Network (CNN) method, to classify bone age utilizing the RSNA dataset and its associated images.This tailored model is adept at recognizing patterns with a newly assembled dataset of regionspecific images, excelling in categorizing diverse bone types.The efficacy of ResNet50 is affirmed through extensive 5-fold crossvalidation, where it outperforms sophisticated models like VGG16 and Xception, attaining outstanding performance metrics with an accuracy of 96.46%, precision of 96.408%, recall of 96.450%, F-score of 96.475%, and specificity of 96.726%.The results of this research carry substantial implications for improving the precise classification of bone age.